DeepSeek’s rise shows why China’s top AI talent is skipping Silicon Valley – Rest of World

This post was originally published on this site.

At the end of his internship at Nvidia in 2023, Zizheng Pan, a young artificial-intelligence researcher from China, faced a pivotal decision: stay in Silicon Valley with the world’s leading chip designers or return home to join DeepSeek, then a little-known startup in eastern China.

Pan chose DeepSeek without much hesitation, Zhiding Yu, a senior research scientist at Nvidia and Pan’s mentor during the internship, recalled on X last month. “I am still very much impressed,” Yu wrote, adding that cases like Pan’s are increasingly common. “Many of our best talents come from China, and these talents don’t have to succeed only in a U.S. company.”

Less than two years after Pan joined DeepSeek, the company catapulted to global fame when it released two AI models that were so advanced, and so much cheaper to build, that the news wiped nearly $600 billion off Nvidia’s market value.

Pan’s choice reflects a growing trend among China’s AI elite to reject Silicon Valley jobs for the AI industry in China, which offers lower living costs, proximity to family, and the opportunity to take on significant roles early in their careers, people in China’s tech industry told Rest of World. DeepSeek filled its ranks with young graduates and interns from elite Chinese universities, such as Tsinghua University and Peking University.

DeepSeek is an outlier in China’s AI industry, as it is fully funded by founder Liang Wenfeng’s trading firm, High-Flyer. The young, passionate tech workers behind DeepSeek are working to catch up with Silicon Valley tech giants, despite the U.S. barring China from accessing advanced chips. 

“What has surprised me is many Chinese students are not that interested in full-time jobs in America.”

“[DeepSeek] highlights the strength of China’s AI talent pool, supported by a large number of highly capable and skilled software engineers,” Angela Zhang, a professor at the University of Southern California who studies tech regulations in China, told Rest of World. “I believe this talent advantage positions China strongly for the next phase of AI development.” 

Nearly half the world’s top AI researchers completed their undergraduate studies in China, according to a 2023 report on global AI talent published by Chicago-based think tank MacroPolo. Chinese universities, state-backed labs, and research arms of American tech giants, such as the Beijing-based Microsoft Research Asia, have helped groom a large group of local researchers.  

For example, Junxiao Song, a core contributor to DeepSeek’s latest R1 model, studied automation at Zhejiang University before obtaining a Ph.D. in electronic and computer engineering at the Hong Kong University of Science and Technology in 2015, according to his Ph.D. adviser Daniel Palomar. Song was persistent and “very mathematically oriented,” Palomar told Rest of World.

When Palomar posted about Song’s work with DeepSeek on LinkedIn, another former student commented that Song used to have the nickname dashi (great master). “Somehow, [DeepSeek] managed to get the best of the best,” Palomar said.  

American companies hire Chinese interns with strong engineering or data-processing capabilities to work on AI projects, either remotely or in their Silicon Valley offices, a Chinese AI researcher at a leading U.S. tech company told Rest of World. “Chinese students do very solid work,” said the researcher, who asked to remain anonymous because he was not authorized to speak to the media. 

But when given full-time offers, many of them have chosen to go back to China, he said. “What has surprised me is many Chinese students are not that interested in full-time jobs in America,” the researcher said. Worries over anti-immigration policies have also deterred some Chinese engineers from moving to the U.S. in recent years. 

Although earlier generations of elite Chinese tech workers preferred Silicon Valley jobs for higher salaries and a chance to work alongside the world’s top innovators, a growing share of young AI engineers are choosing to stay home. There are also more opportunities for them as China’s domestic AI industry expands, with tech giants like Alibaba, and also startups such as StepFun, Minimax, and 01.AI. 

“Somehow, [DeepSeek] managed to get the best of the best.”

DeepSeek has a unique way of wooing talent. The firm pays staff more than ByteDance, according to a recent report from Chinese tech outlet 36Kr. And unlike many Chinese tech firms that foster internal competition and make engineers work grueling hours, Liang told 36Kr in a July 2024 interview that he lets employees find their own tasks and access computing power freely. 

“We are working on the most difficult problems, so we are attractive to them,” he said. 

In the same interview, Liang said making research open-source gives employees a stronger sense of pride and boosts the company’s reputation. Over the past few weeks, some DeepSeek researchers have gained tens of thousands of followers on X, as they discussed research methods and shared their excitement. 

“Beyond words right now,” DeepSeek researcher Deli Chen wrote on X on January 24, as the R1 model ranked number one on a popular chatbot leaderboard. “All I know is we keep pushing forward to make open-source AGI a reality for everyone,” said Chen, who graduated with a master’s degree from Peking University in 2021.

“This moment is absolutely phenomenal to me,” Pan, the former Nvidia intern, wrote two days later. 

Yu Zhou, a professor at Vassar College who has studied the evolution of China’s high-tech industry, told Rest of World the enthusiasm of DeepSeek’s young researchers reminded her of what she had observed at the first internet startups in Beijing in the early 2000s. At the time, graduates from China’s top universities were inspired by the likes of Google and Microsoft, and ended up creating a tech industry at home with less money and fewer top engineers, she said. 

“America thinks China’s trying to unseat America, but the truth is that young people were inspired by new technology developments such as OpenAI,” said Zhou. 

Today’s AI entrepreneurs in China have no choice but to grapple with a shortage of the most powerful Nvidia chips, she said. “When you don’t have resources, all you have is your brain power.”